The capability to incorporate new data sources and to benefit from emerging technologies such as smart meters, web technologies and user experience platforms is critically important if official statistics are to remain relevant. These capabilities require new types of skills and competences that are not all part of the traditional skill set of official statisticians.

While taking care of the production of high-quality statistics, national statistical offices also need to take care of professional statisticians; they have to be encouraged and supported in redefining themselves and in acquiring the new competences required by evolving technologies and practices.

For statisticians to adopt a flexible and pro-active data-driven approach in response to new sources of information, the following features, skills and competences are needed:

• mathematics and statistics (e.g. statistical modeling, machine learning);
• programming and database (e.g. scripting language, SQL);
• domain knowledge;
• soft skills (e.g. team work, creativity);
• communication, storytelling and visualization skills.

All of these data science skills are seldom found in one employee. Instead, statistical offices are introducing collaborative and multidisciplinary data science teams composed of experts with different competences, profiles and backgrounds.

In order to adapt to the evolving professional and technical demands, the necessary competences and skills can be acquired by training current personnel; by collaborating with universities and other networks; or recruiting individual data scientists.

For successful recruitment, statistical offices need personnel planning. This means knowing to what extent data science experts are needed and what kind of positions will be available in the future. Recruiting is not only about employer brand or being appealing as an employer, but ensuring that tasks correspond to the competences of new recruits and that they are able to reach their full potential. This also helps in retaining skilled personnel by ensuring that professional expectations are met.

Data science-driven work demands more than upgrading data science skills in statistical offices. It is just as important to change the working methods and organizational culture. Management and leadership have a crucial role in this respect. Innovation in organizational culture can be supported by:

• acknowledging the interest and potential of current staff e.g. by providing the chance to get familiar with new datasets and try new methods;
• encouraging staff to experiment: allowing staff time to develop new ideas, and accepting occasional failure as a necessary step towards success;
• “letting go of the past” through the ability and willingness to change for the benefit of better statistics.

Together, these actions will help to foster a working environment where professionals, each with their unique experience and identity, love to work together and make good statistics even better.

The issue of how statistical organizations can find, acquire and develop the new generation of statisticians, including concrete follow-up work, were discussed by the 65th Plenary Session of the Conference of European Statisticians (19–21 June 2017). For further information, see: https://www.unece.org/index.php?id=43851#/